advancing the education of our workforce with...
TRANSCRIPT
Advancing the Education of our Workforce with XSEDE
Steven I. Gordon [email protected]
Slides are here
• http://hpcuniversity.org/trainingMaterials/222/
Opportunities and Challenges
• Workforce needs in computational science • Changing how we teach • Barriers to program implementation • Competencies in computational science and
data science • Example Programs • XSEDE Education Program services
Preparing Students
• Need for a workforce which understands both modeling and simulation principles and applications of models and data analysis at large scale – Requirements for high fidelity models of complex systems – Managing and understand large datasets – data science – Applications across a wide range of science, social science,
and increasingly humanities
Crucial Tools for Manufacturing
• At Ford, HPC …allows us to build an environment that continuously improves the product development process, speeds up time-to-market and lowers costs.
• The ongoing use of modeling and simulation resulted in new packaging and product design that propelled the brand to a leading market position over a several-year period.
Ford EcoBoost Technology
Durable coffee package for P&G
Will Pringles Fly?
High Speed Conveying Create Vortices Shedding… …’Rocking Chips’ NOT GOOD!
Brain Cell Linker Dependence Shown by XSEDE Supercomputer Simulations
• Neuroscientists at Stony Brook U. teamed with computational biophysicists at Florida State U. and found that the function of a key brain cell receptor depends on a short polypeptide segment, which they call a linker, to function. – Parkinson's disease, Alzheimer's disease, and a number of
psychiatric disorders are associated with malfunctions of this brain receptor, called the NMDA (N-methyl-D-aspartate) receptor.
– The researchers performed molecular dynamics simulations of the 300,000-atom NMDA receptor system on Stampede.
• Funded by the National Institutes of Health • Published in Nature Neuroscience.
7
8
This protein model represents an NMDA receptor, which juts halfway out of the surface of cells of the nervous system that include the brain and spinal cord. It relays signals between nerve cells. Researchers found that a mechanical coupling was needed between the clamshell-like region at the top of the protein that accepts a neurotransmitter and the channel in the middle that opens for electrical and chemical signals to flow in and out of the brain cell.
Image credit: Lonnie Wollmuth, Stony Brook University
"The resources that are present in the Stampede and XSEDE
facilities go far beyond what can be obtained in one single lab, or
even an institution. It's a very important piece of the
computational infrastructure that's going to be required to
push the frontier of computational research.”
Huan-Xiang Zhou, Florida State University
Myriad of Other Examples
• Behavior of new and existing materials at multiple scales • Climate change and its potential social and economic impacts • Concentration of environmental contaminants and their
impacts on ecosystems and human health • Genetic markers and disease • Analysis of huge datasets
– Market and customer behavior – Genomic data
Changing How We Teach
• Getting students actively involved in learning – Reducing traditional lectures – Increasing inquiry-based learning
• Ideally suited to instruction in computational science – Students need technical and analytical skills to
create and test models and analyze data – Students enhance “soft” skills in teamwork and
written and oral communication
Benefits to Students
• Inquiry-based learning is more effective than traditional lecture oriented instruction – Students are actively engaged in the learning
process – Students gain deeper insights and have higher
retention rates for the information – Facilitates the integration of information across
academic disciplines – math, science, engineering, computer science
11
Challenges to Changing the Curriculum
• We tend to teach in the way we were taught • Computational science is interdisciplinary
– Faculty workloads fixed on disciplinary responsibilities – Coordination across departments is superficial – Expertise at universities is spotty
• Major time commitments are required to negotiate new programs and develop materials
• Curriculum requirements for related fields leave little room for new electives
• Change is hard
Pathways to Reform
• Integrate computational examples into basic science and math courses
• Create general education courses that introduce simulation and modeling concepts and applications
• Combine those efforts to create formal concentrations, minors, or certificates in computational science
• XSEDE is working with institutions to assist with those activities
What Do Students Need to Know?
• Considerable discussion across many disciplines • Difficulty working from general conceptual ideas to
specific skills and knowledge • Several efforts focused on a competency based
model to arrive at consensus of the essential knowledge base
• Competencies reviewed by both academic and non-academic experts
• See http://hpcuniversity.org/educators/competencies/
14
Ohio Minor Program Example
• Undergraduate minor program – 4-6 courses – Varies based on major
• Faculty defined competencies for all students
• Reviewed by business advisory committee
• Program started in Autumn 2007 • Agreements to share students at
distance, instructional modules, revenues, and teaching responsibilities
Competencies for Undergraduate Minor
Simulation and Modeling Programming and Algorithms
Differential Equations and Discrete Dynamical Systems Numerical Methods Optimization Parallel Programming Scientific Visualization One discipline specific course Capstone Research/Internship Experience Discipline Oriented Courses
Adopting the Competencies
• Do not need to incorporate all possible competencies – Adjust based on target student population – Adjust based on available resources
• Does not mean new courses in all areas – Integrate materials in existing courses
Example Competencies Simulation and Modeling
• Explain the role of modeling in science and engineering • Analyze modeling and simulation in computational science • Create a conceptual model • Examine various mathematical representations of functions • Analyze issues in accuracy and precision • Understand discrete and difference-based computer models • Demonstrate computational programming utilizing a higher level language or
modeling tool (e.g. Maple, MATLAB, Mathematica, Python, other) • Assess computational models • Build event-based models • Complete a team-based, real-world model project • Demonstrate technical communication skills
.
Detailed Descriptors
Explain the role of modeling in science and engineering Descriptors: Discuss the importance of modeling to science and engineering Discuss the history and need for modeling Discuss the cost effectiveness of modeling Discuss the time-effect of modeling (e.g. the ability to predict the weather) Define the terms associated with modeling to science and engineering List questions that would check/validate model results Describe future trends and issues in science and engineering Identify specific industry related examples of modeling in engineering (e.g., Battelle; P&G,
material science, manufacturing, bioscience, etc.) Discuss application across various industries (e.g., economics, health, etc.)
Topic Course Credit Hours Terms offered
Required/ Elective
Prerequisites Calculus MATH 1151.xx 5 Au, Sp
MATH 1152.xx or Math 1172 5 Au, Sp
Core Courses
Simulation and Modeling (Choose one of these courses)
MATH 1157 3 Sp Required
CSE 2021 3 Sp ISE 5100 3 Au, Sp ME 5372 3 Au
MATSCEN 4321 3 Au
Programming and Algorithms (Choose one of these courses)
CSE 1222 3 Au, Sp Required
CSE 2221 4 Au, Sp, Su
Numerical Methods (Choose one of these courses)
AERO 3581 3 Au Required
CSE 5361 3 Au, Sp ECE 5510 3 Au MATH 3607 3 Sp MATH 5401 3 Sp
MECHENG 2850 3 Au, Sp
Discipline Specific Courses
Capstone Research/Internship Experience (minimum 3 credits)
CIVILEN 4000.01 2 Au, Sp Required
MATH 4998; CHEM 4998 or other approved individualized research credits **
3-5 Au, Sp, May, Su, May + Su/ Au, Sp/Au, Sp
Discipline-specific Computationally oriented Course
CSE 3521 3 Au, Sp
Required
CSE 3341 3 Au, Sp MICRBIO 5161H 3 N/A BMI 5730 3 Sp CHEM 5440 3 Au MATH 5651 3 Sp PHYS 6810 4 Sp LING 5801 3 Au LING 5802 3 Sp ECON 4050 3 Au, Sp ECON 5001 3 Au, Sp GEOG 5221 3 Au PSYCH 5608 3 Sp
Elective: Choose at least one course from the following (3 credits total required)
Differential Equation and Discrete Dynamical Course
MATH 2255 3 Su, Au, Wi, Sp Elective
MATH 2415 3 Su, Au, Wi, Sp
MATH 2568 3 Su, Au, Wi, Sp
Parallel Programming CSE 5441 3 Au
Elective
Scientific Visualization CSE 5544 1-5 Su, Au, Wi, Sp
Elective
CEG (CIVILEN) 4760 3 Sp E
Data Science Competencies
• Developed by multi-disciplinary group of researchers and practitioners
• Provided as guidance in the development of new programs
• Basic competencies • Advanced competencies
Data Analytics Minor - University of Mary WashingtonTotal credits: 23
Required MATH 220 Introduction to StatisticsCourses MATH 200 Linear algebra
CPSC220 Computer Science 1 Programming and AlgorithmsCPSC419 Data miningCPSC420 Modeling and Simulation
One of these electives CPSC230 Computer Science II Data structuresBUAD 400 Analytics Application Development
One of these electives BUAD 403 Foundations and Applications of Data Analytics CPSC 425 Parallel Processing
Flexibility in Implementation
• Adapt existing courses by adding computationally oriented modules
• Discipline oriented courses dependent on existing faculty expertise and interests
• Different subsets of required and optional competencies tied to major, required math, and example projects
23
XSEDE Education Program
• Campus visits to discuss options • Assistance in curriculum development • Developing Faculty Expertise • Sharing educational materials
Campus Visit
• Half-day to full day visit to campus – Meetings with faculty interested in curriculum changes – Meetings with administrators to explain the need for
changes – Presentation to faculty, students, and administrators – Possible program review exercise
• Target outcome – local committee to work on curriculum change
Developing Faculty Expertise
• Faculty professional development workshops – Two to six day workshops on a variety of topics
• Computational thinking • Computational science education in science and
engineering domains
– Focus on local/regional audiences to reduce travel costs
– Subsidies for faculty to travel to workshops at other sites
26
Collaborative Online Courses
• Courses funded by XSEDE and Blue Waters Projects
• Prepared lectures, quizzes, and exercises online
• Collaborating faculty at local institutions create a local course number and supervise their students
• Use of XSEDE of Blue Waters education allocations
Collaborative Courses
• Applications of Parallel Computers led by University of California Berkeley – XSEDE funding for this year pending tentative spring 2017
• Introduction to HPC led by Kaust – Autumn 2016 sponsored by Blue Waters
• Algorithmic Techniques for Scalable Many-core Computing led by University of Illinois – Autumn 2016 sponsored by Blue Waters
• If interested, email me
Sources for Educational Materials
• Lists of reviewed training and education materials – http://hpcuniversity.org/resources/search/ – http://sighpceducation.acm.org/resources.html
• List of materials and simple modeling tool examples – https://www.osc.edu/~sgordon/workshop/materi
als
29
Ties to Other Community Project
• SIGHPC Education Chapter – http://sighpceducation.acm.org – Webinars on education and training topics – Blog and forum to discuss related issues
• Journal of Computational Science Education – http://jocse.org – Peer reviewed articles on computational science
education applications
How to Get Started
• Review the available materials with the faculty on your campus
• Request a campus visit • Contacts
– Steve Gordon – [email protected] – Kate Cahill – [email protected]
Questions?